Problem Statement

QCSubmit

Currently submitting molecules to the public QCArchive has involved the use of scripts hosted at this github repo . While this method seems to be stable we aim to expand its capabilities while automating the procedure, this will be particularly helpful in the case of the bespoke workflow, but will also impact on the generation of datasets for general force field fitting.

Scope

Must have:

• A workflow, the process to filter and prepare the molecules should be easy to configure and control given a large number of settings available. This should be controllable through the API, settings files and a CLI. Spec for how we imagine this to look can be found here .

• state enumeration stereoisomers/tautomers

• fragmentation

• cmiles identifiers on all submissions

• WBO on all submissions

• filtering based on simple molecule properties (element type size/weight)

• Be easy to serialise to file

• A convenient way to collect the results as well

Nice to have:

• Ability to ensure that calculations can be reused where possible

Not in scope:

Timeline

Milestones and deadlines

Reference materials

Decisions and Problems

1. All OFFTK methods will use OpenEye for consistency until we implement a way to retrieve which backend toolkit performed the requested function.